When something is fully fit for purpose, should you fix it? Most people would say ‘no’. The history of software development seems to be littered with ‘yes’. Take wordprocessors, for example. Back in the day there was WordStar. WordStar was a bit clunky, and required you to use the * key to call up the menus. It didn’t do many of the things you might reasonably want to do, but it was a lot better than an electronic typewriter. After WordStar’s star had really set, WordPerfect and Word came along. On MS-DOS, both of them suffered from not being WYSIWYG, at least, not while you were editing. By suffered, of course, I mean that they were not completely intuitive for formatting letters. You could (and I did) use WordPerfect as a kind of low-grade Desk Top Publishing application, but it was clearly not quite optimised for that.
On the other hand, if you wanted to write a novel, the MS-DOS WordPerfect was pretty much perfect. It didn’t try to distract you with features you didn’t need, didn’t start trying to correct your text while you were writing it, and didn’t involve you in endless font and format choices.
Both Word and WordPerfect moved to Windows versions, and Word got picked up on the Apple Mac. After that, they ‘developed’. By developed I mean that they got more and more complex, offered more and more features, combined some of the attributes of a spreadsheet, a desk top publishing system, a project management system, a reference library, a database manager, and pretty much everything else the programmers thought might be wanted.